This invention relates to optical storage devices typically to replace magnetic disks in digital data processing equipment. Herein usually called the light drive, it typically is characterized by high-speed parallel access to stored data and is a medium size/power/weight device. The light drive uses an optical storage medium to retain data. A photoreceptive material (PRM) retains both wavelength and relative brightness levels of light to encode digital data.
FIG. 1 shows schematically a conceptual diagram of the light drive. A spinning disk coated with PRM provides the data storage. Fixed read and write head arrays provide the electro-optical conversion to the PRM. The light drive includes a data formatter and host interface, both being digital circuits. A power supply is included also.
The Light Drive typically comprises four major subsystems, namely an input/output subsystem, a write subsystem, a read subsystem, and a storage subsystem. The input/output subsystem provides the data interface between the host computer and the light drive, and provides the interface to the read and write subsystems. The input/output subsystem includes the high-speed host data interface and the data formatter/multiplexer, made up of digital circuitry.
The write subsystem provides the electronic data-to-optical conversion for writing data onto the PRM, and includes a number of LED arrays or other high-density light emitting arrays. The write heads use both multiple color arrays and multiple output levels to encode optical data onto the PRM.
The read subsystem provides the optical-to-electronic data conversion for reading data from the PRM, and includes a number of high density CCD arrays or other high density detector arrays. The detector arrays are color filtered for each band of wavelength detection, and provide multiple level output.
The storage subsystem provides the optical data storage, including the spinning disk coated with PRM, and any required uninterruptable power supply (UPS). The storage subsystem includes the disk, motor, and physical mounting assembly for the light drive.
The input/output subsystem requires both a host data interface and a data formatter. The host data interface provides a high-speed mass storage interface. This interface typically is an industry standard interface in the one to five gigabytes per second range. There is no pressing requirement to exceed the useful data rate of the host.
Current enabling technologies for the host data interface include Fibre Channel, both in single and multiple links. The host data interfaces typically are identical to those for high-speed magnetic disk arrays, with no unique problems for the light drive.
The data formatter buffers data for disassembly and reassembly to/from the read/write system. The data formatter multiplexes data into parallel channels, corresponding to the parallel (concentric) tracks on the disk. The formatter partitions data bits between wavelength and level prior to optical encoding, and provides the massive interconnect to the read and write subsystems. Additionally, the data formatter may need to provide a refresh capability for destructive-read PRM (see PRM section).
Current high speed digital design methodology and materials are sufficient to realize the data formatter. The multiplexing functions can be provided by current field programmable gate array (FPGA) and application specific integrated circuit (ASIC) technology. Buffering functions can be provided by current static read only memory (SROM) technology.
The write subsystem typically comprises a large number of parallel tracks (T) at high-density spacing (S). A number of multiple monochromatic light sources (C) are provided at controlled brightness levels (L bits, or 2L levels). The combination of all of these preferably provide a terabyte or more of storage (Txc3x97Sxc3x97Cxc3x97L greater than 1 terabyte). The multiple light sources preferably have sufficiently narrow light spectra for separation at L levels. The light sources also preferably have sufficient uniformity to accurately modulate 2L levels, and a high switching speed to allow a write throughput of at least about 100 megabytes per second. The write head should have sufficient light conversion efficiency to meet reasonable power requirements and should have reasonable component costs.
Write head-enabling technologies include light emitting diode (LED) linear arrays. These were originally developed by OKI Electric Industry for laser quality printers. OKI Data is currently (November 1999) in production with LED array printers at both 300 and 600 dots per inch (dpi). In July 1998, OKI developed a new fabrication process to make 1200 dpi LED arrays. The fabrication process currently limits a single array to 1200 dpi resolution. It may be possible to achieve 2400 dpi using dual offset write heads. The multi-color support of the process is not known. The 1200 dpi array is about 6% efficient, which is sufficient for this application. The approximately 3% corrected uniformity of this array allows up to four bit modulation (16 levels). OKI suggests usage at 44 kHz or more. The switching speed may not be limited by the LED technology but more by power considerations.
Typical multicolor LED technology limits the Light Drive to five colors maximum. These are yellow (570 nm), Red (660 nm), and three infrared (850, 880, and 940 nm). It is currently not known whether other colors besides red can be fabricated into high-density arrays using the OKI process.
Another potential write head technology is organic LED (OLED). Emerging OLED technology may allow higher density or more colors than current LED technology. Other flat panel display technologies currently under development may be applicable to a write head array, including field emission display (FED), gas plasma, and liquid crystal display (LCD).
The read subsystem typically comprises a large number of parallel tracks (T) at high density spacing (S), and some multiple filtered light detectors (C) with sufficient detection levels (typically L bits, or 2L levels). The combination of all of these should provide a terabyte or more of storage (Txc3x97Sxc3x97Cxc3x97L greater than 1 terabyte). The multiple filtered detectors should have sufficiently narrow light spectra for separation at L levels, with sufficient uniformity and low noise to accurately detect 2L levels. The detectors should have a high enough sensing speed to allow a read throughput of at least about 100 megabytes per second. The read head should have sufficient sensitivity to match PRM output, and should have reasonable component costs.
The read head enabling technologies typically include charge coupled device (CCD) linear arrays, contact image sensors (CIS), and a unique LED read/write head. The CCD-based read head typically uses multiple color filtered monochrome arrays. The CCDs have a wide dynamic range (typically about 10,000:1) and high sensitivity (typically about 0.6 xcexcJ/cm2 for 100% output). Current CCDs are high density with pixels typically sized down to about 6.5 xcexcm square. Current CCDs, however, are not high-speed devices. The fastest available CCD array provides about 13.6 kHz line rate. Thus, current CCD technology makes the light drive read-side speed somewhat limited. Also, the pitch of current CCDs typically does not match the LED-based write head.
CIS devices are similar to CCDs and are available in about the same size and pitch as LED arrays. CISs are rather slow, perhaps too slow for best results with the light drive. OKI developed an LED read/write head for a low cost FAX (facsimile) application in April 1997. This device has similar read characteristics to a CIS. Other potential read head technologies include photodiode arrays, phototransistor arrays, and higher speed CCD technologies.
The storage subsystem typically comprises a PRM, spinning disk substrate, and optionally a UPS. The PRM retains both relative wavelengths of light and relative light levels for data storage.
The spinning disk provides a stable rotating surface for the PRM. A prototype has been designed using currently available LED arrays for the write head and CCDs for the read head, with operation typically at up to about 36 RPM for a 20-inch diameter disk. This is reasonable using current technology.
A UPS may be required to maintain the PRM memory during a power failure. One kilowatt of sustain power is considered reasonable. Power outages of several hours should be tolerated. Thus, the UPS, if needed, may require up to several kilowatt-hours of capacity, which is currently available. The UPS is not required if the PRM is self-stable.
The PRM stores data by retaining the relative wavelengths and luminance levels of light in its material. The PRM is xe2x80x9cexposedxe2x80x9d during a write operation to multiple colors of light each modulated to one of M levels. Each data cell contains several bits layered in a number of colors, C. Since each color is modulated to M levels, each data cell contains C log2(M) levels. The system typically comprises five colors at 16 levels each, for a total of 5 log2(16)=20 bits per cell.
The PRM is stimulated to emit for a read operation. The PRM either emits the same spectrum as was written, or a shifted spectrum which retains the relative wavelength information. Relative luminance is also maintained at each recovered wavelength. In either case, the brightness level information at each color is recovered without interference to allow full recovery of recorded data.
The PRM provides unlimited read/write cycles at sufficient switching speed to support at least about 100 megabytes per second throughput. It provides non-volatile storage that typically is stable for multiple years with up to 1 kW steady state power applied.
The PRM typically has a write cycle in the xcexcJ/cm2 range for reasonable light drive power consumption. It also has a low sustain power, in the nanowatt per cell range, for cell retention; and should have reasonable emission power.
The PRM should be capable of addressable emission, because the material cannot continuously emit at the required detection level without requiring unreasonable power requirements. Thus, some mechanism typically is provided to stimulate emission.
The PRM can have either destructive or non-destructive read characteristics. With destructive read, the data typically is refreshed after each read by looping read data back to the write subsystem. This typically involves a slight complexity increase in the light drive electronics. Data is erased using dummy reads. With non-destructive read, an erase mechanism is provided prior to write operations.
PRM should be applied to and adhere to the disk surface. The PRM preferably should not require coherent (laser) sources, and should have reasonable material cost.
The identity of the preferred PRM is not yet certain. It appears that various organic compounds will be preferred.
Organics are of considerable interest for optical storage media as they offer a wide range of features that are attractive for both permanent and erasable/rewritable memories. These attractive features for organic storage media include high stability, low melting points, low thermal conductivities, a wide variety of spectroscopic transitions, and simpler less expensive manufacturing processes. Mechanisms for optical storage of information in organic materials including polymers involves processes such as ablation, bubble formation, phase changes, color bleaching and photochromism. Generally, information is written into a medium at one wavelength and read out at another wavelength, a different polarization, or through a change in scattering or reflectivity of light.
Further details are covered briefly below on some of the mechanisms with examples of organic materials that are of interest for high-density optical storage media.
Storage of optical information in organic media involves the switching of a bistable molecular state to another state. One important mechanism that is receiving considerable attention today is photochromism where molecules are converted between two forms, X and Y. Ultraviolet light is usually used to convert the X state to the Y state with the Y state readout at a longer, less energetic wavelength. In some cases, a third wavelength can be used to switch the Y state back to the X state. Two examples of photochromic organic compounds that operate in this mode are the dihydroindoligines[1] and 4xe2x80x2 methoxyflavylium perchlorate[2]. These molecular switching systems operate in the write-lock-read-unlock-erase mode. A photochromic material, 1,3,3-trimethylindolkino-6xe2x80x2-nitrobenzopyrylospiran in polystyrene, appears promising for a 3-D optical memory[3]. Excitation at 442 nm converts the molecule to an isomer with strong absorption at 612 nm, which cannot be used for readout since absorption at that wavelength returns the molecules to the original isomer. However, the two isomers exhibit different refractive indices, which can be used for readout.
Two-step excitation at two different wavelengths or two-photon excitation of photochromism also is used for optical memories. In this mode, information may be recorded at the intersection of two light beams. An example of a two-photon approach, a naphthopyran derivative is used with an irradiation wavelength of 405 nm and erasure at 334 nm. Readout is based on a color change[4]. A number of organometallic charge-transfer compounds have been found useful for optical information storage using reflectivity changes for readout[5,6].
Biological materials in combination with polymers have been of interest for optically switchable devices. An example is the use of bacteriorhodopsin in polyvinyl alcohol[7]. Images are written in yellow light, which can be erased either by blue light which switches the yellow molecules to purple, or by yellow light, that converts the purple molecules to yellow. Dye-doped liquid-crystalline materials also are finding use for long-term optical storage media.[8,9]
Bleaching utilizes a material where the recording mechanism involves a colored/colorless transition. One example involves 6xe2x80x2-nitroindolinospiropyan. Radiation at 580 nm changes the colored form to colorless, which is reversed with radiation at 380 nm[10]. Two-photon polymer bleaching mechanisms also appear promising for 3-D memories[11].
The above information on optical memories based on organic material was intended to be brief with emphasis on recent developments. The future for optical storage media based on organic and polymeric materials appears very promising. Information on the common holographic storage media can be found in a summary by Kampf and Mergel[12].
[1] F. Pina, M. J. Melo, M. Maestri, R. Ballardini, and V. Balzani, xe2x80x9cPhotochromism of 4xe2x80x2-Methoxyflavylium Perchlorate. A xe2x80x9cWrite-Lock-Read-Unlock-Erasexe2x80x9d Molecular Switching System,xe2x80x9d J. Am. Chem. Soc. 119, 5556 (1997).
[2] C. Weber, F. Rustemeyer, and H. Durr, xe2x80x9cA Light-Driven Switch Based on Photochromic Dihydroindolizinesxe2x80x9d Adv. Mater. 10, 1348 (1998).
[3] A. Toriumi, J. M. Herrmann, and S. Kawata, xe2x80x9cNondestructive Readout of a Three-Dimensional Photochromic Optical Memory with a Near-infrared Differential Phase-Contrast Microscope,xe2x80x9d Opt. Lett. 22, 555 (1997).
[4] M. Uchida and M. Irze, xe2x80x9cTwo-Photon Photochromism of a Naphthopyran Derivative,xe2x80x9d J. Am. Chem. Soc. 115, 6442 (1993).
[5] R. C. Hoffman and R. S. Potember, xe2x80x9cOrganometallic Materials for Erasable Optical Storage,xe2x80x9d App7. Opt. 28, 1417 (1989).
[6] C. Y. Liu and A. J. Bard, xe2x80x9cOptoelectronic Properties and Memories Based on Organic Single-Crystal Thin Films,xe2x80x9d Acc. Chem. Res. 32, 235 (1999).
[7] Z. Chen, A. Lewis, H. Takei, and I. Nebenzahl, xe2x80x9cBacteriorhodopsin Oriented in Polyvinyl Alcohol Films as an Erasable Optical Storage Medium,xe2x80x9d Appl. Opt. 30, 5188 (1991).
[8] S. Hvilsted, F. Andruzzi, and P. S. Ramaniujam, xe2x80x9cSide-Chain Liquid-Crystalline Polyesters for Optical Information Storage,xe2x80x9d Opt. Lett. 17, 1234 (1992).
[9] F. Simoni, O. Francascangeli, Y. Reznikor, and S. Slussarenko, xe2x80x9cDye-Doped Liquid Crystals as High-Resolution Recording Media,xe2x80x9d Opt. Lett. 22, 549 (1997).
[10] C. Lenoble and R. S. Becken, xe2x80x9cPhotophysics, Photochemistry, Kinetrics, and Mechanisms of the Photochromism of 6xe2x80x2-Nitroindolinopiropyran,xe2x80x9d J. Phys. Chem. 90, 62 (1986).
[11] D. Day and M. Gu, xe2x80x9cEffects of Refractive-Index Mismatch on Three-Dimensional Optical Data-Storage Density in a Two-Photon Bleaching Polymer,xe2x80x9d App. Opt. 37, 3299 (1998).
[12] G. Kampf and D. Mergel, xe2x80x9cInformation Storage Materials. Optical,xe2x80x9d in Kirk-Othmer Encyclopedia of Chemical Technology, Fourth Edition, Volume 14, p. 277, J. I. Kroschwitz and M. Howe-Grant, Editors, John Wiley and Sons. (1995).